By W. L. Rathje

Most archaeologists have an inferiority complex - and who can blame them? Researchers who study today’s societies spend their time asking "live" residents questions about what is going on and "waste" very little time looking at smelly garbage for answers. But for archaeologists, who study "dead" civilizations, garbage is about all that is left to answer their questions.

What’s an archaeologist to do? Realize something that all can tossers, haulers, and all others who deal with discards on a daily basis know: There is as much or more reality in garbage than in answers to interview-surveys or questionnaires.

That reality came home to roost at the Garbage Project when, in 1986, the Census Bureau asked the Garbage Project to help solve one of its chronic problems: undercounting. The bureau, by most accounts, has done a near-perfect job of counting mainstream Americans, but its enumerators have been missing a large share of adult males among undocumented aliens and residents of urban ghettos (especially in areas where the presence of an adult male in a household can affect welfare checks) - exactly those who don’t want to be counted.

The Census Bureau came to the Garbage Project to see whether data derived from refuse could reliably be used to check the bureau’s counts in problem neighborhoods. The project’s job was to find out if it was possible to roughly reconstruct a neighborhood’s population by age and sex simply on the basis of what that neighborhood threw away.

To an archaeologist, of course, the idea of using garbage to reconstruct population characteristics was not in the least bizarre. For most, population estimates are just the number of identified dwellings multiplied by some "magic number" that is an educated guess at the average number of residents per household.

The Garbage Project’s first concern was to develop an analysis format that completely protected anonymity - no names or addresses were recorded and garbage was analyzed only by whole neighborhoods. Once anonymity was assured, the study for the Census Bureau became a search for a magic number - a multiplier that, when applied to quantities of particular kinds of garbage, would yield accurate population estimates. We began by analyzing a mass of computerized evidence acquired during studies of food consumption and garbage production from about 200 households where all residents were known by age and sex. The data were compiled by actually sorting, counting, weighing, and recording the garbage from these households (with each household’s permission) over a five-week period.

The study quickly determined that if you multiplied a constant (the magic number) by the weight of all the garbage collected from a set number of households over a specific period of time - minus the weight of yardwastes, which varied greatly between inner city, suburbs, and so on - you could accurately estimate the number of people who lived in those households. It turned out that the categories "total solid waste" and "plastic" had the best predictive power.

The equation based on total solid waste is, however, less universally reliable than the one for plastic. This is because children are responsible for less garbage overall that adults. Yet plastic is another story. During any given period of time, every man, woman, and child seems to generate about the same amount of plastic - usually in the form of many small items. Plastic is America’s great garbage equalizer!

For a neighborhood of 100 households, the projected population estimate derived from the average of total solid waste and plastic results, applied to one week’s worth of garbage, was accurate to within ±2.5%. That is considerably better than the Census Bureau can do in many places.

Overall population is only one demographic characteristic, however. What about estimates of age and sex? This turned out to be a trickier proposition. The easiest subpopulation to identify is infants. Disposable diapers are a convenient marker, and infants go through so many of them in a week that they are an ideal item for establishing correlations.

Infant-size disposable diapers are exclusively worn by infants. But when it comes to distinguishing between men and women, or middle-aged and elderly adults, there are fewer exclusives. For example, disposable razors may indicate the presence of men, but don’t necessarily do so, since women use them as well. Those items that are "near" exclusives, such as fat cigars and men’s jockey shorts, are discarded very infrequently and therefore have little predictive value.

It took a considerable amount of work, but the Garbage Project staff eventually came up with an equation for estimating the number of children in a population. The estimate is based on the average rate of discarded toys and toy packages and children’s clothes and packages. It was also possible to derive equations for estimating the proportion of adult women in a population, based on the number of discarded female-hygiene products, cosmetics, and women’s clothing items.

Finding serviceable material correlates for estimating the proportion of adult men - he Census Bureau’s ultimate objective - proved more elusive. Men are not totally invisible in garbage, but garbage is an unreliable indicator of their live-in presence. Women may drink and eat like men. They smoke cigarettes. They sometimes wear men’s clothing and cologne. Even the presence of male contraceptive packaging is at best uncertain evidence of a long-term male household member.

In the end, the best way to get a figure for the number of adult men in a given neighborhood turns out to be a backdoor procedure. First, find the total neighborhood population. Next, subtract the estimates for infants, children, and adult women from the total population estimate. The result is an estimate of the adult male population, and it has an accuracy of better than ±10% - and the Census Bureau has been accused of being off in its counts of males by 40% or more in minority neighborhoods.

A garbage "census" would clearly be a usable snapshot of low-income neighborhoods. As it happened, however, the Garbage Project never got the chance. In 1988, the director of the Census Bureau’s Center for Survey Methods Research decided that, from a public-relations standpoint, "it is risky for the government to hire someone to analyze garbage." A year later, the bureau announced its decision not to adjust the 1990 census to compensate for the expected undercount, a decision that has stayed in place in 2000. Recently, accounts have emerged that the rich are getting harder to count as well - their rate of return of census forms has plummeted, and gated communities have not been easy for census-takers to swing open.

The Garbage Project stands ready for the year 2010!

Contributing author W. L. Rathje is director of the Garbage Project.

MSW
September/October, 2000